Which automation tool should we use? It is one of the first questions people ask when they start thinking about workflow automation, and it is almost always the wrong question to start with. The tool is the last decision, not the first. But it is a decision that needs to be made well, because the wrong choice at this stage creates real costs: migration overhead, per-task billing that grows unexpectedly, or technical complexity the team cannot maintain.
What follows is the practical breakdown of the three tools that cover the vast majority of small service business automation needs: Zapier, Make (formerly Integromat), and n8n. The comparison is based on direct experience building and maintaining automations in all three, across a range of business types and workflow profiles.
The three tools in a paragraph each
Zapier is the most accessible automation tool available. It is visual, no-code, and connects to over 7,000 applications. It was designed for non-technical users who want to connect two or three apps without writing any logic or understanding what an API is. The interface is deliberately simple: a trigger event in one app causes one or more actions in other apps. For straightforward, linear workflows, this is exactly what you need and Zapier delivers it reliably. The constraint is the pricing model: Zapier charges per task, where each action in an automation counts as one task. A business running ten automations at moderate volume can spend CHF 300 to CHF 600 per month. At scale, this cost grows with usage in a way that Make and n8n do not.
Make (formerly Integromat) offers significantly more visual power than Zapier through a flowchart-based builder that handles multi-step logic, conditional branching, loops, and error handling more naturally. The visual representation of a Make scenario resembles the actual workflow it represents, which makes complex automations easier to build, read, and debug. Pricing is based on operations (roughly equivalent to tasks in Zapier), but the structure tends to produce lower costs at equivalent automation volume. The integration library is smaller than Zapier's at around 1,500 native integrations, but the HTTP module allows connection to any system with an API, which covers most real-world needs. The learning curve is steeper than Zapier. A non-technical team member can use Zapier within an hour; Make typically takes a day or two to become comfortable with.
n8n is an open-source automation tool with a fundamentally different economic model. Self-hosted, it carries near-zero running cost: you pay only for the server it runs on, typically CHF 10 to CHF 20 per month on a basic cloud instance. There is no per-task billing, no operations limit, no monthly ceiling. The trade-off is technical requirement: n8n expects users who can read JSON, understand data structures, diagnose integration errors, and maintain a server. For a business with a technical co-founder or an operations person with developer experience, n8n is the most cost-effective and most flexible tool available. For a business without that profile, it is not the right choice regardless of the cost savings.
Comparison at a glance
Which tool fits which business profile
Zapier is the right choice when: you need to connect two or three well-known apps with simple, linear logic. You have a clear use case, you want it working this week, and you do not have someone on the team who will enjoy debugging a more complex tool. The budget is not the primary constraint. You are building two to five automations and do not plan to scale significantly beyond that. In this profile, Zapier's simplicity and reliability are genuine advantages. The per-task cost is not a concern at low volume.
Make is the right choice when: you have five to fifteen workflows to automate, at least some of which involve conditional logic, multi-step sequences, or data transformations. You have someone on the team who is comfortable with visual tools and can troubleshoot when something breaks. You want lower running costs than Zapier without the technical overhead of self-hosting. Data residency in the EU is required or preferred. Make is the tool MEIKAI uses as a default for most Swiss SMB clients: strong enough for complex workflows, cost-effective at scale, and accessible enough that a non-developer team member can maintain it after implementation.
n8n is the right choice when: the business has a technical person who can own the infrastructure. You have high workflow volume, strict data residency requirements, or complex integration needs that go beyond what Zapier and Make handle well. You want no running cost beyond server fees and no per-task limits. MEIKAI deploys n8n for clients who have these characteristics, particularly practices with strict Swiss data residency requirements who want their automation infrastructure entirely on Swiss-hosted servers.
The tool is not the bottleneck
The most common error in automation projects is treating tool selection as the first decision. The tool is the last decision, because the tool should be determined by the workflow requirements, not the other way around.
Starting with a tool preference and then bending your workflows to fit it creates three problems. First, you may pick a tool that handles simple workflows well but breaks down when you encounter the complex cases that always appear in real-world implementations. Second, you may overbuild: using n8n for workflows that Zapier would handle perfectly, adding technical complexity and maintenance overhead where none is needed. Third, you may underbuild: using Zapier for a workflow that requires conditional logic and error handling, and spending CHF 40 per month on a brittle automation that fails every third run.
Starting with the wrong automation is more expensive than starting with the wrong tool. A bad tool choice creates migration overhead. A bad workflow choice means you have automated something that should not have been automated first, while the high-value workflow continues to consume time. Get the workflow selection right first; the tool selection follows naturally from the workflow requirements.
What MEIKAI uses and why
MEIKAI is not committed to any single tool. Every Clarity Scan produces a tool recommendation as part of the implementation plan, based on the specific workflow requirements and the technical profile of the team that will own the automations after delivery.
In practice, Make is the default recommendation for most Swiss SMB clients. The reasons are consistent: it handles the multi-step, conditional workflows that most small practices need; EU data processing options address the nFADP/FADP compliance requirements that Swiss businesses face; the cost model is predictable at scale; and it is accessible enough that a non-developer can maintain it after the initial implementation is complete.
n8n is recommended for clients with high workflow volume, strict Swiss data residency requirements, or an internal technical resource who will actively develop and maintain the automation layer. The zero per-task cost model is a meaningful advantage for practices running twenty or more automations at significant volume.
Zapier is recommended when the client's specific use cases are well-served by its simplicity, when time to deployment is the primary constraint, or when the client wants to manage a small number of automations independently without relying on MEIKAI for ongoing support. It is also the right choice for clients who are testing the automation approach for the first time and want to validate the return before committing to a more capable tool.
Custom code is used where none of the three tools handles the integration reliably, or where a specific API interaction is complex enough that a purpose-built script is more maintainable than an automation scenario. This is uncommon in small business implementations but does arise, particularly with legacy practice management software that has limited or poorly documented API support.